Current infrared detector systems incorporate arrays of large numbers of discrete, highly sensitive detector elements, the outputs of which are connectable to sophisticated processing circuitry. By rapidly analyzing the pattern and sequence of detector element excitations, the processing circuitry can identify and monitor sources of infrared radiation. A contemporary subarray of detectors may contain 256 detectors on a side, or a total of 65,536 detectors. The size of each square detector is approximately 0.009 cm on a side with 0.00127 cm spacing between detectors. Such a subarray would therefore be 2.601 cm on a side. The subarray may, in turn, be joined to form an array that connects twenty-five million or more detectors. Considerable difficulties are presented in normalizing the output of each detector.
The response of an infrared detector and its associated electronic processing channel is linear with an output measured in volts per watt of absorbed infrared radiation. A DC offset is associated with each detector and signal processing channel and is defined as fixed pattern noise. The detector response and DC offset are unique for each channel, i.e. they are different for each channel but fall within a given range of values. Satisfactory methods exist to eliminate fixed pattern noise, however the removal of variations in channel response, or gain among elements within an array is typically accomplished with a separate specialized computer.
Gain normalization to a high level is required by the sophisticated algorithms which analyze the detector array outputs. The variation in gain from one detector channel to another increases the dynamic range requirements for the output amplifier when detector channels are multiplexed into a common output amplifier. The dynamic range of the analog to-digital converter and the signal processing hardware must likewise have sufficient dynamic range to handle the multiplexed data.
Increasing the dynamic range requirements of the output amplifier, analog-to-digital converter, and signal processing hardware results in an increase in system weight and cost. Therefore, it is desirable to maintain the best possible uniformity in gain among all detector channels. Gain normalization commonly consists of multiplication of the signal by a correction factor after analog-to-digital conversion. This is commonly accomplished with a separate specialized computer which is required to have sufficient memory to store a correction factor for each channel and must also have dedicated multiplication hardware. The memory and power requirements become severe for a five million or more element detector array.
The use of such a specialized computer to perform gain normalization also increases the size and weight of the detector system. This is particularly crucial in spacecraft applications where weight, volume, and power considerations are paramount. As such, although the prior art has recognized the need for gain normalization for infrared detector arrays, the proposed solutions have to date been ineffective in providing a satisfactory remedy.